Re: [PR c++/82836] Fixe testcase

[Rainer Orth]

Hi Nathan,

> The 82836 testcase fell out of creduce.  In c++17 mode it fails horribly
> with missing return errors.
>
> Applying this fix, so it's valid in c++17.  It still ICEs (in both 14 and
> 17 modes) with the 82836 fix removed.

unfortunately, the testcase has several more problems:

* It fails on 32-bit Solaris/SPARC:

Excess errors:
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:12:3: error: 
expected unqualified-id before '__int128'
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:13:3: error: 
'__float128' does not name a type; did you mean 'float'?
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:158:31: error: 
'operator new' takes type 'size_t' ('unsigned int') as first parameter 
[-fpermissive]
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:159:33: error: 
'operator new' takes type 'size_t' ('unsigned int') as first parameter 
[-fpermissive]
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:285:37: error: 
template argument 1 is invalid
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:288:46: error: 
template argument 1 is invalid

* Likewise on 64-bit Solaris/SPARC:

Excess errors:
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:13:3: error: 
'__float128' does not name a type; did you mean 'float'?

* And again on 32-bit x86 (both Solaris and Linux):

Excess errors:
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:12:3: error: 
expected unqualified-id before '__int128'
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:158:31: error: 
'operator new' takes type 'size_t' ('unsigned int') as first parameter 
[-fpermissive]
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:159:33: error: 
'operator new' takes type 'size_t' ('unsigned int') as first parameter 
[-fpermissive]
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:285:37: error: 
template argument 1 is invalid
/vol/gcc/src/hg/trunk/local/gcc/testsuite/g++.dg/pr82836.C:288:46: error: 
template argument 1 is invalid

All of those are fixed as follows.  Tested with the appropriate runtest
invocations on i386-pc-solaris2.11, sparc-sun-solaris2.11, and
x86_64-pc-linux-gnu.

Ok for mainline, or would you rather make the use of __int128 and
__float128 conditional?

Thanks.
        Rainer

-- 
-----------------------------------------------------------------------------
Rainer Orth, Center for Biotechnology, Bielefeld University


2017-11-19  Rainer Orth  <r...@cebitec.uni-bielefeld.de>

        gcc/testsuite:
        * g++.dg/pr82836.C: Require int128, __float128 support.
        Add __float128 options.
        (size_t): Define using __SIZE_TYPE__.


# HG changeset patch
# Parent  a9e82327ce2d2f2e3fd4a800f29cf023cbc455e2
Add g++.dg/pr82836.C requirements

diff --git a/gcc/testsuite/g++.dg/pr82836.C b/gcc/testsuite/g++.dg/pr82836.C
--- a/gcc/testsuite/g++.dg/pr82836.C
+++ b/gcc/testsuite/g++.dg/pr82836.C
@@ -1,12 +1,15 @@
 // PR c++/82836
 // { dg-options "-std=c++17" }
+// { dg-require-effective-target int128 }
+// { dg-require-effective-target __float128 }
 // { dg-additional-options "-Wno-pedantic -Wno-return-type" }
+// { dg-add-options __float128 }
 // We were resetting DECL_ASSEMBLER_NAME when processing pragma weak,
 // breaking C++'s mangling alias hash table.  That hash table needs to
 // be tickled in just the right way to hit the problem.
 
 namespace std {
-typedef long unsigned size_t;
+typedef __SIZE_TYPE__ size_t;
 inline namespace __cxx11 {}
   double abs() {return 0;}
   __int128 abs(__int128 ) {return 0;}